Freezing shoulder clinical staging prediction method, equipment and program product

Abstract

The invention relates to the field of intelligent medical treatment, in particular to a frozen shoulder clinical staging prediction method and device and a program product. Comprising the steps that S1, passive motion range data of shoulder joints of a to-be-detected person is obtained, and the passive motion range comprises inward rotation, outward rotation, stretching, buckling and outward stretching; s2, performing first classification based on the passive activity data to obtain classification results of normal to-be-tested persons and abnormal to-be-tested persons; and S3, transmitting the passive activity data of the abnormal person to be detected to a freezing shoulder mathematical model for classification to obtain classification results of an adhesion early stage, a gradual freezing stage, a freezing stage and an unfreezing stage, wherein the freezing shoulder mathematical model is obtained by calculating a mapping relationship between the passive activity and image features of different stages. Different periods of the frozen shoulder can be distinguished through the passive activity range, accurate prevention and treatment are facilitated, and good clinical value is achieved.

Description

Technical Field

[0001] This application relates to the field of intelligent medicine, and particularly to a prediction method, device, program product and computer-readable storage medium for the clinical staging of frozen shoulder. Background Art

[0002] Frozen shoulder (FS) is one of the common pathological conditions in the shoulder joint, characterized by gradually increasing shoulder pain and limited range of motion. Clinically, FS is also known as adhesive capsulitis of the shoulder joint and belongs to a type of shoulder inflammation and fibrosis lesion. Epidemiological studies have shown that the average duration of FS is 2 - 3 years. Up to 40% of patients may have persistent symptoms and limited mobility for more than 3 years, and 15% of patients may result in permanent disability. It has been reported that the occurrence of FS is closely related to diabetes, Dupuytren's contracture, shoulder injury, body mass index (≥25 kg / m2) and age (56 - 70 years). Although FS is less common than diseases such as subacromial impingement syndrome, calcifying tendinitis and rotator cuff tear, it is sometimes difficult to distinguish based on its clinical manifestations. In addition, the treatment methods for FS are also different at different stages. In the past, the clinical diagnosis of FS was usually based on medical history and physical examination, and there was no clear key value to evaluate the condition of FS. Therefore, exploring some imaging manifestations for the diagnosis and staging of FS is of great significance for further clinical treatment. In addition, frozen shoulder usually develops slowly and can be divided into four stages: pre-adhesive stage, sub-freezing stage, freezing stage, and thawing stage. Each stage can last for several months, but clinically, it is more common to classify the sub-freezing stage, freezing stage, and thawing stage. Although patients in the pre-adhesive stage will have severe pain at the end of the joint range of motion, pain at rest, obvious pain at night, and sleep will be affected, and timely intervention after the early detection of the symptoms of frozen shoulder can avoid the deterioration of the disease. Summary of the Invention

[0003] In view of the above problems, the present invention provides a prediction method for the clinical staging of frozen shoulder, specifically including:

[0004] S1. Obtain the passive range of motion data of the shoulder joint of the person to be tested. The passive range of motion includes internal rotation, external rotation, extension, flexion, and abduction;

[0005] S2. Perform a first classification based on the passive range of motion data to obtain the classification results of normal and abnormal persons to be tested; the first classification is performed by comparing with a first preset threshold; the first preset threshold is the normal boundary value of the passive range of motion;

[0006] S3. Input the passive range of motion data of the abnormal person to be tested into the frozen shoulder mathematical model for classification to obtain the classification results of the pre-adhesive stage, sub-freezing stage, freezing stage, and thawing stage. The frozen shoulder mathematical model is obtained by calculating the mapping relationship between the passive range of motion and the imaging features of different stages.

[0007] The passive range of motion of the abnormal person to be tested is input into the frozen shoulder mathematical model. When the passive range of motion does not satisfy the mapping relationship of the frozen shoulder mathematical model, it is predicted as the pre-adhesion stage;

[0008] Optionally, the construction process of the frozen shoulder mathematical model is as follows:

[0009] Obtain high-frequency ultrasound images and passive range of motion during the three stages of the frozen shoulder: the amyotrophic stage, the frozen stage, and the thawing stage;

[0010] Extract features from the high-frequency ultrasound images to obtain image features;

[0011] Based on the three stages of the amyotrophic stage, the frozen stage, and the thawing stage, calculate the first threshold, the second threshold, and the third threshold of the image features respectively;

[0012] Calculate the mapping relationship between the passive activity degrees in the three stages and the first threshold, the second threshold, and the third threshold of the image features to obtain the frozen shoulder mathematical model;

[0013] Optionally, the image features include one or more of the following: the thickness of the inferior joint capsule, the shoulder joint distance, and the joint capsule thickness.

[0014] The image feature is the thickness of the inferior joint capsule. Based on the images of the three stages of the amyotrophic stage, the frozen stage, and the thawing stage, calculate the three-stage thresholds of the thickness of the inferior joint capsule, and calculate the mapping relationship between the three-stage thresholds of the thickness of the inferior joint capsule and the passive activity degrees in the three stages to obtain the mathematical model of the frozen inferior joint shoulder; classify through the mathematical model of the frozen inferior joint shoulder to obtain the classification results of the pre-adhesion stage, the amyotrophic stage, the frozen stage, and the thawing stage;

[0015] Optionally, the image feature is the shoulder joint distance. Based on the images of the three stages of the amyotrophic stage, the frozen stage, and the thawing stage, calculate the three-stage thresholds of the shoulder joint distance, and calculate the mapping relationship between the three-stage thresholds of the shoulder joint distance and the passive activity degrees in the three stages to obtain the mathematical model of the frozen shoulder with joint distance; classify through the mathematical model of the frozen shoulder with joint distance to obtain the classification results of the pre-adhesion stage, the amyotrophic stage, the frozen stage, and the thawing stage;

[0016] Optionally, the image feature is the joint capsule thickness. Based on the images of the three stages of the amyotrophic stage, the frozen stage, and the thawing stage, calculate the three-stage thresholds of the joint capsule thickness, and calculate the mapping relationship between the three-stage thresholds of the joint capsule thickness and the passive activity degrees in the three stages to obtain the mathematical model of the frozen joint shoulder; classify through the mathematical model of the frozen joint shoulder to obtain the classification results of the pre-adhesion stage, the amyotrophic stage, the frozen stage, and the thawing stage.

[0017] The S3 is replaced with: inputting the passive range of motion of the abnormal subject to be measured into the second adhesive capsulitis model for classification to obtain the classification results of the pre-adhesion stage, the amyotrophic stage, the frozen stage, and the thawing stage. The second adhesive capsulitis model includes L sub-models, where L is a natural number greater than 1. Each sub-model is obtained by calculating the mapping relationship between the passive range of motion and the imaging features of different stages. The imaging features of each sub-model are different, and the prediction results of the sub-models are fused to obtain the final prediction result;

[0018] Optionally, the sub-models include the lower joint adhesive capsulitis mathematical model, the joint distance adhesive capsulitis mathematical model, and the joint adhesive capsulitis mathematical model.

[0019] The S3 is replaced with: obtaining the shoulder imaging data of the abnormal subject to be measured; extracting features from the shoulder imaging data to obtain imaging features; performing second classification based on the imaging features to obtain the classification results of the pre-adhesion stage, the amyotrophic stage, the frozen stage, and the thawing stage;

[0020] Optionally, the second classification is performed by comparing the imaging feature values with the second preset thresholds to obtain the classification results, where the second preset thresholds include the normal imaging feature threshold, the amyotrophic stage imaging feature threshold, the frozen stage imaging feature threshold, and the thawing stage imaging feature threshold.

[0021] The second classification is performed by an imaging classification model trained by imaging features and labels to obtain the classification results;

[0022] Optionally, the training process of the imaging classification model includes:

[0023] obtaining the imaging data and labels of the abnormal subject with passive range of motion to be measured and extracting features to obtain imaging features;

[0024] inputting the imaging features into a neural network model for training until the loss function of the model remains unchanged to obtain the imaging classification model;

[0025] Optionally, the labels include normal, amyotrophic stage, frozen stage, and thawing stage; when the classification result of the imaging classification model is normal, it is determined as the pre-adhesion stage;

[0026] Optionally, the neural network includes one or more of the following: random forest, support vector machine, decision tree, XGBoost, AdaBoost, convolutional neural network, and residual network.

[0027] The method further includes clinical data. Based on the clinical data and the adhesive capsulitis model, classification is performed to obtain the classification results of the pre-adhesion stage, the amyotrophic stage, the frozen stage, and the thawing stage. The clinical data includes one or more of the following: Constant-Murley score, visual analogue scale.

[0028] The present invention provides another clinical staging method for frozen shoulder, including: replacing S1 with: obtaining the imaging data of the subject to be tested; replacing S2 with: extracting features from the imaging data to obtain imaging features, including the inferior capsular thickness (ICT) and the glenohumeral distance (GHD); replacing S3 with: performing three-classification based on the imaging features to obtain the classification results of the amyotrophic stage, frozen stage, and thawing stage.

[0029] The imaging features further include the anterior capsular thickness (ACT);

[0030] Optionally, the imaging includes high-frequency ultrasound imaging and MR imaging;

[0031] Optionally, the inferior capsular thickness (ICT) and the glenohumeral distance (GHD) are extracted from the high-frequency ultrasound imaging, and the anterior capsular thickness (ACT) is extracted from the MR imaging;

[0032] Optionally, the three-classification is obtained by comparing the imaging features with the preset thresholds of the amyotrophic stage, frozen stage, and thawing stage;

[0033] Optionally, the three-classification is replaced with: classifying through a classification model trained with the imaging feature data of the amyotrophic stage, frozen stage, and thawing stage to obtain the classification results;

[0034] Optionally, the method further includes clinical data, obtaining clinical data, and classifying based on the clinical data and imaging features to obtain the classification results. The clinical data includes one or more of the following: Constant-Murley score, visual analogue score.

[0035] The purpose of the present invention is to provide a computer program product, which includes a computer program or instruction, and the computer program or instruction is executed by a processor to implement the above-mentioned prediction method for clinical staging of frozen shoulder.

[0036] The purpose of the present invention is to provide a computer device, which includes a memory, a processor, and a computer program or instruction stored on the memory, and the computer program or instruction is executed by the processor to implement the above-mentioned prediction method for clinical staging of frozen shoulder.

[0037] The purpose of the present invention is to provide a computer-readable storage medium, on which a computer program or instruction is stored, and the computer program or instruction is executed by a processor to implement the above-mentioned prediction method for clinical staging of frozen shoulder.

[0038] Advantages of the present invention:

[0039] 1. During different stages of frozen shoulder, the range of shoulder movement is restricted, and as the condition progresses, the degree of restriction gradually worsens. In the prior art, three-stage classification of frozen shoulder is performed through imaging, but the pre-adhesion stage between the normal state and the frozen stage of frozen shoulder is ignored. And during the pre-adhesion stage, there is no adhesion during arthroscopic examination (imaging cannot directly diagnose the pre-adhesion stage), but the patient will experience pain during rest, especially obvious pain at night, which affects sleep. In early clinical diagnosis, it is often suspected of subacromial impingement. Therefore, the present invention uses passive range of motion to stage the pre-adhesion stage, frozen stage, frozen period, and thawing period of frozen shoulder, avoiding misdiagnosis in the pre-adhesion stage of frozen shoulder, providing a more complete diagnostic method for frozen shoulder, and having good clinical value.

[0040] 2. Predictive classification of the four stages of frozen shoulder is performed by combining passive range of motion with imaging features. Based on the passive range of motion of the subject to be tested, it is determined whether there is an abnormality. For subjects with abnormalities, further predictive judgments of the pre-adhesion stage, frozen stage, frozen period, and thawing period of frozen shoulder are made through imaging features. The advantages of imaging features for the three-stage classification of frozen shoulder are fully utilized to assist the passive range of motion in identifying different stages of frozen shoulder, reducing the probability of misdiagnosis, and improving the pertinence and effectiveness of prevention and treatment.

[0041] 3. The present invention extracts ICT, GHD, and ACT imaging features from high-frequency ultrasound images and MR images respectively for classification of the frozen stage, frozen period, and thawing period, which can effectively identify frozen shoulder at different stages and achieve corresponding treatment and timely treatment. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for description in the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those skilled in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0043] Figure 1 It is a schematic flowchart of the prediction method for the clinical staging of frozen shoulder provided by the embodiment of the present invention;

[0044] Figure 2 It is a schematic diagram of the prediction system for the clinical staging of frozen shoulder provided by the embodiment of the present invention;

[0045] Figure 3 It is a schematic diagram of the prediction device for the clinical staging of frozen shoulder provided by the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0046] In order to enable those skilled in the art to better understand the solution of the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention.

[0047] In some of the processes described in the specification, claims, and the above-mentioned drawings of the present invention, a plurality of operations appear in a specific order. However, it should be clearly understood that these operations may not be executed in the order in which they appear herein or may be executed in parallel. The serial numbers of the operations, such as S101, S102, etc., are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions such as "first" and "second" herein are used to distinguish different messages, devices, modules, etc., and do not represent a sequence, nor do they limit that "first" and "second" are of different types.

[0048] Figure 1 The schematic diagram of the prediction method for the clinical staging of frozen shoulder provided by the embodiments of the present invention specifically includes:

[0049] S101: Obtain the passive range of motion data of the shoulder joint of the person to be tested. The passive range of motion includes internal rotation, external rotation, extension, flexion, and abduction;

[0050] In one embodiment, frozen shoulder is also known as adhesive capsulitis. The shoulder joint is as stiff as if it were frozen, restricting the range of motion of the shoulder joint, so it is vividly called frozen shoulder. It is a relatively common disease clinically.

[0051] Internal rotation: For the 90° examination, the arm on the examined side automatically extends backward, passes behind the body, and the hand can touch the opposite rib or scapula. For those with pain, passive examination can be performed.

[0052] External rotation: 60°, the hand can actively touch the occiput and the opposite ear. For those with pain, passive external rotation examination can be performed.

[0053] Flexion 180°, abduction 180°, extension 50°. The end points of flexion and abduction are the same point, which is called coupled movement. The range of motion of the shoulder joint is evaluated by the angle.

[0054] In a specific embodiment, according to the medical records of the patient, the passive range of motion is used to judge the nature of the joint motion end sensation and determine whether there is any abnormal structure restricting the joint motion. The examination content mainly includes the movement of several angles such as internal rotation, external rotation, extension, flexion, and abduction.

[0055] S102: Perform a first classification based on the passive range of motion data to obtain the classification results of normal and abnormal persons to be tested; the first classification is performed by comparing with a first preset threshold;

[0056] The first preset threshold is the normal boundary value of the passive range of motion;

[0057] In one embodiment, when the subject under test performs internal rotation, external rotation, extension, flexion, and abduction, the angles that can be completed are used to determine whether the completed angles reach the completed angles or angle ranges of normal people. When not satisfied, it is determined as abnormal.

[0058] S103: Input the passive range of motion data of the abnormal subject under test into the adhesive capsulitis mathematical model for classification to obtain the classification results of the pre-adhesion stage, the progressive freezing stage, the frozen stage, and the thawing stage. The adhesive capsulitis mathematical model is obtained by calculating the mapping relationship between the passive range of motion and the imaging features of different stages.

[0059] In one embodiment, when the passive range of motion of the abnormal subject under test is input into the adhesive capsulitis mathematical model, if the passive range of motion does not satisfy the mapping relationship of the adhesive capsulitis mathematical model, it is predicted as the pre-adhesion stage.

[0060] In one embodiment, the construction process of the adhesive capsulitis mathematical model is as follows:

[0061] Obtain high-frequency ultrasound images and passive range of motion of the progressive freezing stage, frozen stage, and thawing stage of adhesive capsulitis;

[0062] Extract imaging features from the high-frequency ultrasound images;

[0063] Based on the progressive freezing stage, frozen stage, and thawing stage, calculate the first threshold, second threshold, and third threshold of the imaging features respectively;

[0064] Calculate the mapping relationship between the passive activity of the three stages and the first threshold, second threshold, and third threshold of the imaging features to obtain the adhesive capsulitis mathematical model.

[0065] In one embodiment, the imaging features include one or more of the following: the thickness of the inferior joint capsule, the shoulder joint distance, and the joint capsule thickness.

[0066] In one embodiment, by performing high-frequency ultrasound imaging and MR on the subject with abnormal passive range of motion, extract the inferior joint capsule thickness ICT and the shoulder joint distance GHD from the high-frequency imaging, and extract the joint capsule thickness ACT from the magnetic resonance imaging.

[0067] In one embodiment, the imaging feature is the thickness of the inferior joint capsule. Based on the images of the progressive freezing stage, frozen stage, and thawing stage, calculate the three-stage thresholds of the thickness of the inferior joint capsule, and calculate the mapping relationship between the three-stage thresholds of the thickness of the inferior joint capsule and the passive activity of the three stages to obtain the inferior joint adhesive capsulitis mathematical model; classify through the inferior joint adhesive capsulitis mathematical model to obtain the classification results of the pre-adhesion stage, the progressive freezing stage, the frozen stage, and the thawing stage.

[0068] In one embodiment, the image feature is the shoulder joint distance, and the three-stage thresholds of the shoulder joint distance are calculated based on the three-stage images of the gradual freezing period, the freezing period, and the thawing period, respectively. The mapping relationship between the three-stage thresholds of the shoulder joint distance and the three-stage passive activity is calculated to obtain a mathematical model of the frozen shoulder of the joint distance; the mathematical model of the frozen shoulder of the joint distance is used to classify and obtain the classification results of the early adhesion period, the gradual freezing period, the freezing period, and the thawing period;

[0069] In one embodiment, the image feature is the thickness of the joint capsule. The three-stage thresholds of the joint capsule thickness are calculated based on the three-stage images of the gradual freezing period, the freezing period, and the thawing period, respectively. The mapping relationship between the three-stage thresholds of the joint capsule thickness and the three-stage passive activity is calculated to obtain a mathematical model of the frozen shoulder. The mathematical model of the frozen shoulder is used for classification to obtain the classification results of the pre-adhesion period, the gradual freezing period, the freezing period, and the thawing period.

[0070] In one embodiment, the image features are the thickness of the lower joint capsule, the shoulder joint distance, and the joint capsule thickness. The three-stage thresholds of the image features are calculated based on the three-stage images of the gradual freezing period, the freezing period, and the thawing period. The mapping relationship between the three-stage thresholds of the thickness of the lower joint capsule, the shoulder joint distance, and the joint capsule thickness and the three-stage passive activity is calculated to obtain a mathematical model of frozen shoulder.

[0071] In one embodiment, the abnormal passive range of motion is input into the frozen shoulder mathematical model. When the passive range of motion is classified into a certain category through a mapping relationship, it is determined to be the corresponding classification category. When the passive range of motion is not classified into a certain category through a mapping relationship, it is determined to be the early stage of adhesion.

[0072] In one embodiment, S103 is replaced by: inputting the passive range of motion of the abnormal subject into the second frozen shoulder model for classification to obtain the classification results of the early adhesion stage, gradual freezing stage, freezing stage, and thawing stage. The second frozen shoulder model includes L sub-models, L is a natural number greater than 1, and each sub-model is obtained by calculating the mapping relationship between the passive range of motion and the imaging features of different stages. The imaging features of each sub-model are different, and the prediction results of the sub-models are fused to obtain the final prediction result.

[0073] Optionally, the sub-models include a lower joint frozen shoulder mathematical model, a joint distance frozen shoulder mathematical model, and a joint frozen shoulder mathematical model.

[0074] In one embodiment, the abnormal passive range of motion is input into the second adhesive capsulitis of the shoulder mathematical model, wherein the passive range of motion is input into the sub-models in parallel, and the mapping relationship of the passive range of motion is judged in each sub-model. When the classification category is obtained through the mapping relationship of the passive range of motion, it is determined as the belonging classification category. When the classification category cannot be obtained through the mapping relationship of the passive range of motion, it is determined as the early stage of adhesion. When the classification results of the sub-models are the same, the classification result is output. When the classification results of the sub-models are different, the final classification category is obtained through Bayesian voting decision on the output results of each sub-model.

[0075] In one embodiment, S103 is replaced by: obtaining the shoulder image data of the abnormal person to be measured; extracting features from the shoulder image data to obtain image features; performing a second classification based on the image features to obtain the classification results of the early stage of adhesion, the anesthetic stage, the frozen stage, and the thawing stage.

[0076] Optionally, the second classification is performed by comparing the image feature values with the second preset thresholds to obtain the classification results, where the second preset thresholds include the normal image feature threshold, the anesthetic stage image feature threshold, the frozen stage image feature threshold, and the thawing stage image feature threshold.

[0077] In one embodiment, the second classification is performed by an image classification model trained by image features and labels to obtain the classification results.

[0078] In one embodiment, the training process of the image classification model includes:

[0079] Obtaining the image data and labels of the abnormal passive range of motion person to be measured and extracting features to obtain image features;

[0080] Inputting the image features into the neural network model for training until the loss function of the model remains unchanged to obtain the image classification model.

[0081] Optionally, the labels include normal, anesthetic stage, frozen stage, and thawing stage; when the classification result of the image classification model is normal, it is determined as the early stage of adhesion.

[0082] In one embodiment, the neural network includes one or more of the following: random forest, support vector machine, decision tree, XGBoost, AdaBoost, convolutional neural network, and residual network.

[0083] In one embodiment, the method further includes clinical data, and the classification results of the early stage of adhesion, the anesthetic stage, the frozen stage, and the thawing stage are obtained based on the clinical data and the adhesive capsulitis of the shoulder model. The clinical data includes one or more of the following: Constant-Murley score, visual analogue scale.

[0084] In one embodiment, a method for classifying the three stages of frozen shoulder is provided, and S101 is replaced with: obtaining the image data of the subject to be measured;

[0085] S102 is replaced with: extracting features from the image data to obtain image features, including the inferior capsular thickness ICT and the capsular distance GHD;

[0086] S103 is replaced with: performing three-class classification based on the image features to obtain the classification results of the amyotrophic stage, the frozen stage, and the thawing stage.

[0087] In one embodiment, the image features further include the capsular thickness ACT;

[0088] Optionally, the images include high-frequency ultrasound images and MR images;

[0089] Optionally, the inferior capsular thickness ICT and the capsular distance GHD are extracted from the high-frequency ultrasound images, and the capsular thickness ACT is extracted from the MR images.

[0090] In one embodiment, the three-class classification is obtained by comparing the image features with the preset thresholds of the amyotrophic stage, the frozen stage, and the thawing stage.

[0091] In one embodiment, the three-class classification is replaced with: performing classification through a classification model trained with the image feature data of the amyotrophic stage, the frozen stage, and the thawing stage to obtain the classification results.

[0092] In one embodiment, the method further includes clinical data, obtaining clinical data, and performing classification based on the clinical data and the image features to obtain the classification results. The clinical data includes one or more of the following: Constant-Murley score, visual analogue score.

[0093] By concurrently performing stage determination based on clinical data and stage determination based on image features, when the two determination results are the same, the classification results are output; when the two determination results are different, the final classification results are obtained through a Bayesian voting mechanism.

[0094] In another embodiment, the clinical data and the image features are input into a classification model to be trained for training to obtain a trained multi-modal classification model, and classification is performed based on the multi-modal classification model to obtain the classification results.

[0095] In a specific embodiment, the medical data of FS patients (from January 2021 to February 2022, n = 100) were sorted out by the present invention. According to the FS stage, the patients were divided into stage I group (n = 30), stage II group (n = 35) and stage III group (n = 35). The FS stage was divided according to the report of "Adhesive Bursitis: A Review of Current Treatment". (1) Stage I FS: There is mild pain in the shoulder, mainly manifested as soreness, cold pain, stabbing pain and dull pain. The severity of the symptoms varies during the day and at night, being lighter during the day and heavier at night. In severe cases, it may affect sleep, and the functional activity of the shoulder joint is slightly limited or normal. Arthroscopic examination shows extensive fibrous synovitis in the scapular joint, especially in the upper part of the anterior joint capsule, without adhesion contracture of the joint capsule. Pathology: Mild infiltration of inflammatory cells, rich blood supply, bursa hyperplasia inflammation, normal joint capsule tissue. (2) Stage II FS: Severe shoulder pain, especially at night, seriously affecting the patient's sleep, and accompanied by restriction of the overall range of motion. Arthroscopic examination shows a pedunculated hyperplastic synovitis reaction, and partial disappearance of the axillary bursa. Pathology: Rich blood supply, synovial hyperplasia inflammation, a large number of mast cells and chromaffin cells in the synovial tissue, and scarring of the bursa. (3) Stage III FS: Compared with the most severe stage, the shoulder pain is reduced, but the mechanical obstruction of the shoulder is still obvious. Arthroscopic examination shows a synovitis reaction, and complete disappearance of the axillary bursa. Pathology: Dense cell hyperplasia, collagen tissue hyperplasia, and thinning of the synovium.

[0096] Inclusion criteria: (1) Patients who underwent HFU or MR examination; (2) Age ≥ 18 years old; (3) Course of disease ≥ 2 months; (4) Patients who met the diagnostic criteria of FS stage; (5) Normal consciousness level and voluntary participation. Exclusion criteria: (1) Incomplete clinical information; (2) Secondary FS caused by rheumatic diseases; (3) Severe impairment of liver and kidney and other organ functions; (4) Infection; (5) History of surgery in the shoulder joint or adjacent area.

[0097] HFU examination: A Logiq E9 color Doppler ultrasound diagnostic instrument (ML6-15, USA) was purchased from General Electric Company, USA. The patient sat on an adjustable-height swivel chair. The examiner first scanned from the back of the shoulder joint to find the longitudinal axis section of the attachment end of the infraspinatus tendon of the humerus, moved the probe along the longitudinal axis of the infraspinatus tendon of the humerus, showed the posterior glenoid labrum between the back of the humeral head and the glenoid fossa, and measured the humeroglenoid distance. Then, the patient's upper arm was placed on the maximum support, and the probe was placed under the armpit to show and measure the maximum thickness of the glenohumeral joint capsule. During the examination, the shoulder being examined was passively moved, and the boundaries and adhesions of the joint capsule were dynamically observed. HFU-related indicators included the inferior capsule thickness (ICT, measured from the cortical bone of the humerus to the thickest part of the lateral edge of the joint capsule) and the humeroglenoid distance (GHD, the vertical distance between the lateral edge of the glenoid labrum and the humeral head). The HFU examination was performed by an ultrasound doctor with more than 5 years of work experience. Diagnostic criteria for HFU examination of FS: (1) ICT ≥ 3 mm; (2) Reduced echo in the rotator interval with increased blood flow signal.

[0098] MR examination: A superconducting magnetic resonance imaging instrument was purchased from Siemens (Magnetom Avanto 1.5T, Germany). The patient lay supine on the magnetic resonance examination table. The affected shoulder was placed at the center of the magnet, and the upper limb was kept in a standard neutral position. Scanning parameters included axial scanning, oblique sagittal scanning, and oblique coronal scanning. The matrix was set to 256×205, the scanning range was 180×180 mm, and the slice thickness was 3 mm. MR-related indicators included the joint capsule thickness (ACT), the coracohumeral ligament thickness (CHLT), and the rotator interval-joint capsule thickness (RIACT). The MR examination was performed by a radiologist with more than 3 years of work experience. Diagnostic criteria for MR examination of FS: (1) Thickening of the joint capsule; (2) Edema in the rotator interval.

[0099] In a specific embodiment, clinical data collection and analysis were obtained from the patient's medical records, including gender (male and female), age, height, weight, affected shoulder (left and right), Constant-Murley score (CMS), visual analogue score (VAS), and passive range of motion (internal rotation (IR), external rotation (ER), extension (EX), flexion (FL), and abduction (AB)).

[0100] CMS (0-100 points) included four parts: the degree of shoulder joint pain (0-15 points), the ability of daily life and activities of the shoulder joint (0-20 points), the active range of motion of the shoulder joint (0-40 points), and the muscle strength of the shoulder joint (0-25 points). The higher the CMS score, the better the shoulder joint function.

[0101] The VAS score is a commonly used method to evaluate the degree of pain clinically. According to the participants' subjective pain sensations, the participants are guided to mark the corresponding pain points on a long line segment. The VAS score ranges from 0 (no pain) to 10 (severe pain). The higher the VAS score, the more severe the shoulder pain indicates.

[0102] Statistical analysis was performed using SPSS 20.0 (SPSS Inc., USA). Statistical charts were drawn using Graphpad 8.0 (Graphpad Inc., USA). Continuous variables that conformed to a normal distribution were expressed as the mean ± standard deviation. One-way analysis of variance or Welch's test was used for multi-group comparisons. The Bonferroni method or Tamhane's T2 method was used for multiple comparisons between groups. Continuous variables that did not conform to a normal distribution were expressed by the quartile method [M (P25, P75)]. The Kruskal Wallis H test was used for comparisons between groups. Categorical variables were expressed as ratios or constituent ratios, and the χ2 test of the R×C contingency table or the χ2 test of the R×C contingency table was used for comparison. The diagnostic value of each parameter was analyzed by the receiver operating characteristic (ROC) curve, and the values of the area under the ROC curve (AUC), confidence interval (CI), sensitivity, specificity, and optimal cut-off point were obtained. AUC > 0.9: the highest diagnostic value; 0.8 < AUC < 0.9: excellent diagnostic value; 0.5 < AUC < 0.8: low diagnostic value. The significance level was two-sided P<0.05.

[0103] In a specific embodiment, as shown in Table 1, the average ages of the first-phase group (n = 30), the second-phase group (n = 35), and the third-phase group (n = 35) were 53.37±2.83, 52.94±2.39, and 53.97±1.89, respectively. In addition, there were no significant differences among the three groups regarding gender, age, height, weight, and the affected shoulder (Table 1, all P>0.05). It is worth noting that there were obvious differences among the three groups regarding CMS, VAS, and passive range of motion (Table 1, P<0.05).

[0104] As shown in Table 2, it can be seen that there are significant differences among the three groups in terms of ICT (3.01 (2.99, 3.05) vs. 3.44 (3.40, 3.52) vs. 3.19 ± 0.09, P < 0.05) and GHD (3.13 ± 0.18 vs. 2.06 ± 0.12 vs. 2.43 ± 0.23, P < 0.05). On the other hand, there are obvious differences among the three groups in terms of ACT (4.96 ± 0.34 vs. 3.95 ± 0.22 vs. 3.30 ± 0.20; Table 3, P < 0.05) and CHLT (4.15 ± 0.21 vs. 4.08 ± 0.24 vs. 3.98 ± 0.28; Table 3, P < 0.05). However, there are no significant differences among the three groups in terms of RIACT (Table 3, P > 0.05).

[0105] Table 1

[0106]

[0107] Table 2

[0108]

[0109] Table 3

[0110]

[0111] The diagnostic consistency rates of the three diagnostic methods were evaluated (Table 4). After high-frequency ultrasound diagnosis, there were 91 positive cases and 9 negative cases, and the diagnostic consistency rate was 91.00%. After magnetic resonance diagnosis, there were 93 positive cases and 7 negative cases, and the diagnostic consistency rate was 93.00%. Under the combined diagnosis of high-frequency ultrasound and magnetic resonance, there were 96 positive cases and 4 negative cases, and the diagnostic consistency rate was 96.00%.

[0112] Table 4

[0113]

[0114] In a specific embodiment, the research results showed that in the diagnosis of the first stage of FS, ICT (AUC = 0.914), GHD (AUC = 0.999), and ACT (AUC = 1.000) had high diagnostic values (Table 5, P < 0.05). In the diagnosis of the second stage of FS, ICT (AUC = 0.947) and GHD (AUC = 0.974) had high diagnostic values (Table 6, P < 0.05). While in the diagnosis of the third stage of FS, ACT (AUC = 0.989) had high diagnostic value (Table 7, P < 0.05).

[0115] Table 5

[0116]

[0117] Table 6

[0118]

[0119] Table 7

[0120]

[0121] In a specific embodiment, the results showed that there was a negative correlation between ICT and VAS, and a positive correlation between GHD / ACT / CHLT and VAS (Table 8, P < 0.05). As for CMS, there was a negative correlation between ICT / ACT and CMS, and a positive correlation between GHD and CMS (Table 9, P < 0.05).

[0122] Regarding the correlation between HFU / MR indicators and passive range of motion related indicators (Table 10), it can be seen that there was a positive correlation between ICT / ACT and IR, and a negative correlation between ICT / ACT and ER / EX / FL / AB (P < 0.05); there was a negative correlation between GHD and IR, and a positive correlation between GHD and ER / EX / FL / AB (P < 0.05).

[0123] Table 8

[0124]

[0125] Table 9

[0126]

[0127] Table 10

[0128]

[0129] This study evaluated the diagnostic value of HFU and MR-related indicators in the diagnosis of FS staging. The data of the present invention showed that in the first stage of FS, ICT and ACT had high diagnostic value; in the second stage of FS, ICT and GHD had high diagnostic value; and in the third stage of FS, ACT had high diagnostic value. These results indicated that these parameters might become diagnostic indicators for FS staging. In addition, the present invention also conducted a correlation analysis between HFU / MR indicators and VAS / CMS / passive range of motion. The research results showed that ICT was negatively correlated with VAS, while GHD / ACT / CHLT were positively correlated with VAS. For CMS, ICT / ACT were negatively correlated with CMS, while GHD was positively correlated with CMS. In addition, ICT / ACT were positively correlated with IR, while ICT / ACT were negatively correlated with ER / EX / FL / AB. GHD was negatively correlated with IR, while GHD was positively correlated with ER / EX / FL / AB. These research results were discovered for the first time and were helpful for formulating targeted FS treatment measures.

[0130] The disclosed embodiment of the present invention also provides a computer program product or system, including a computer program which, when executed by a processor, implements the steps of the above-mentioned prediction method for the clinical staging of frozen shoulder.

[0131] Figure 2 Schematic diagram of the prediction system for the clinical staging of frozen shoulder provided by the embodiment of the present invention, specifically including:

[0132] Acquisition module: to acquire the passive range of motion data of the shoulder joint of the person to be tested, and the passive range of motion includes internal rotation, external rotation, extension, flexion, and abduction;

[0133] First classification module: to perform a first classification based on the passive range of motion data to obtain the classification results of normal and abnormal persons to be tested; the first classification is performed by comparing with a first preset threshold; the first preset threshold is the normal boundary value of the passive range of motion;

[0134] Second classification module: to input the passive range of motion data of the abnormal person to be tested into the frozen shoulder mathematical model for classification to obtain the classification results of the pre-adhesion stage, the progressive freezing stage, the freezing stage, and the thawing stage, and the frozen shoulder mathematical model is obtained by calculating the mapping relationship between the passive range of motion and the imaging features of different stages.

[0135] Figure 3 Schematic diagram of the prediction device for the clinical staging of frozen shoulder provided by the embodiment of the present invention, specifically including:

[0136] A memory and a processor; the memory is used to store program instructions; the processor is used to call the program instructions, and when the program instructions are executed, it performs any one of the above-mentioned prediction methods for the clinical staging of frozen shoulder.

[0137] The disclosed embodiments of the present invention also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements any of the above-described prediction methods for the clinical staging of frozen shoulder.

[0138] The verification results of this verification embodiment show that assigning inherent weights to indications can improve the performance of this method compared to the default settings. Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the above-described systems, devices, and units can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein. In the several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there can be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of devices or units can be in electrical, mechanical, or other forms. The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment. In addition, in each embodiment of the present invention, the functional units can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-integrated units can be implemented in the form of hardware or in the form of software functional units. Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing relevant hardware through a program, and this program can be stored in a computer-readable storage medium. The storage medium can include: read-only memory (ROM, Read Only Memory), random access memory (RAM, Random Access Memory), magnetic disk, or optical disc, etc.

[0139] Those of ordinary skill in the art can understand that all or part of the steps in implementing the methods of the above embodiments can be completed by instructing relevant hardware through a program, and the program can be stored in a computer-readable storage medium. The above-mentioned storage medium can be read-only memory, magnetic disk, or optical disc, etc.

[0140] The above has introduced in detail a computer device provided by the present invention. For those of ordinary skill in the art, according to the idea of the embodiments of the present invention, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation on the present invention.

Claims

1. A prediction method for the clinical staging of frozen shoulder, characterized in that, Including: S1. Obtain the passive range of motion data of the shoulder joint of the person to be tested. The passive range of motion includes internal rotation, external rotation, extension, flexion, and abduction. S2. Perform a first classification based on the passive range of motion data to obtain the classification results of normal and abnormal persons to be tested. The first classification is performed by comparing with a first preset threshold. S3. Input the passive range of motion data of the abnormal person to be tested into the adhesive capsulitis mathematical model for classification to obtain the classification results of the pre-adhesion stage, the frozen stage, the frozen period, and the thawing period. The adhesive capsulitis mathematical model is obtained by calculating the mapping relationship between the passive range of motion and the imaging features of different stages.

2. The prediction method for the clinical staging of frozen shoulder according to claim 1, characterized in that When the passive range of motion of the abnormal person to be tested is input into the adhesive capsulitis mathematical model and does not satisfy the mapping relationship of the adhesive capsulitis mathematical model, it is predicted as the pre-adhesion stage. Optionally, the construction process of the adhesive capsulitis mathematical model is as follows: Obtain the high-frequency ultrasound images and passive range of motion of the frozen stage, the frozen period, and the thawing period of adhesive capsulitis. Extract features from the high-frequency ultrasound images to obtain imaging features. Based on the frozen stage, the frozen period, and the thawing period, calculate the first threshold, the second threshold, and the third threshold of the imaging features respectively. Calculate the mapping relationship between the passive activity degrees of the three stages and the first threshold, the second threshold, and the third threshold of the imaging features to obtain the adhesive capsulitis mathematical model. Optionally, the imaging features include one or more of the following: the thickness of the inferior joint capsule ICT, the glenohumeral distance GHD, and the thickness of the joint capsule ACT.

3. The prediction method for the clinical staging of frozen shoulder according to claim 2, wherein The imaging feature is the thickness of the inferior joint capsule. Calculate the three-stage thresholds of the thickness of the inferior joint capsule based on the images of the frozen stage, the frozen period, and the thawing period respectively. Calculate the mapping relationship between the three-stage thresholds of the thickness of the inferior joint capsule and the passive activity degrees of the three stages to obtain the inferior joint adhesive capsulitis mathematical model. Classify through the inferior joint adhesive capsulitis mathematical model to obtain the classification results of the pre-adhesion stage, the frozen stage, the frozen period, and the thawing period. Optionally, the imaging feature is the glenohumeral distance. Calculate the three-stage thresholds of the glenohumeral distance based on the images of the frozen stage, the frozen period, and the thawing period respectively. Calculate the mapping relationship between the three-stage thresholds of the glenohumeral distance and the passive activity degrees of the three stages to obtain the glenohumeral distance adhesive capsulitis mathematical model. Classify through the glenohumeral distance adhesive capsulitis mathematical model to obtain the classification results of the pre-adhesion stage, the frozen stage, the frozen period, and the thawing period. Optionally, the imaging feature is the thickness of the joint capsule. Calculate the three-stage thresholds of the thickness of the joint capsule based on the images of the frozen stage, the frozen period, and the thawing period respectively. Calculate the mapping relationship between the three-stage thresholds of the thickness of the joint capsule and the passive activity degrees of the three stages to obtain the joint adhesive capsulitis mathematical model. Classify through the joint adhesive capsulitis mathematical model to obtain the classification results of the pre-adhesion stage, the frozen stage, the frozen period, and the thawing period. Optionally, S3 is replaced with: inputting the passive range of motion of the abnormal subject to be measured into the second adhesive capsulitis of the shoulder model for classification to obtain the classification results of the pre-adhesion stage, the amyotrophic stage, the frozen stage, and the thawing stage. The second adhesive capsulitis of the shoulder model includes L sub-models, where L is a natural number greater than 1. Each sub-model is obtained by calculating the mapping relationship between the passive range of motion and the imaging features of different stages. The imaging features of each sub-model are different, and the prediction results of the sub-models are fused to obtain the final prediction result; Optionally, the sub-models include the lower joint adhesive capsulitis of the shoulder mathematical model, the joint distance adhesive capsulitis of the shoulder mathematical model, and the joint adhesive capsulitis of the shoulder mathematical model.

4. The prediction method for the clinical staging of frozen shoulder according to claim 1, wherein S3 is replaced with: obtaining the shoulder imaging data of the abnormal subject to be measured; extracting features from the shoulder imaging data to obtain imaging features; performing second classification based on the imaging features to obtain the classification results of the pre-adhesion stage, the amyotrophic stage, the frozen stage, and the thawing stage; Optionally, the second classification is performed by comparing the imaging feature values with the second preset thresholds to obtain the classification results, where the second preset thresholds include the normal imaging feature threshold, the amyotrophic stage imaging feature threshold, the frozen stage imaging feature threshold, and the thawing stage imaging feature threshold; Optionally, the second classification is performed by an imaging classification model trained by imaging features and labels to obtain the classification results; Optionally, the training process of the imaging classification model includes: obtaining the imaging data and labels of the abnormal subject with passive range of motion to be measured and performing feature extraction to obtain imaging features; inputting the imaging features into a neural network model for training until the loss function of the model remains unchanged to obtain the imaging classification model; Optionally, the labels include normal, amyotrophic stage, frozen stage, and thawing stage; when the classification result of the imaging classification model is normal, it is determined as the pre-adhesion stage; Optionally, the neural network includes one or more of the following: random forest, support vector machine, decision tree, XGBoost, AdaBoost, convolutional neural network, and residual network.

5. The prediction method for the clinical staging of frozen shoulder according to claim 1, wherein The method further includes clinical data, and classification is performed based on the clinical data and the adhesive capsulitis of the shoulder model to obtain the classification results of the pre-adhesion stage, the amyotrophic stage, the frozen stage, and the thawing stage. The clinical data includes one or more of the following: Constant-Murley score, visual analogue scale score.

6. The prediction method for the clinical stages of frozen shoulder according to claim 1, characterized in that S1 is replaced with: obtaining the imaging data of the subject to be measured; S2 is replaced with: extracting features from the imaging data to obtain imaging features, including the inferior joint capsule thickness ICT and the joint capsule distance GHD; S3 is replaced with: performing three-classification based on the imaging features to obtain the classification results of the amyotrophic stage, the frozen stage, and the thawing stage.

7. The prediction method for the clinical staging of frozen shoulder according to claim 6, wherein, The imaging features further include the articular capsule thickness ACT; Optionally, the imaging includes high-frequency ultrasound imaging and MR imaging; Optionally, the inferior joint capsule thickness ICT and the joint capsule distance GHD are extracted from the high-frequency ultrasound imaging, and the articular capsule thickness ACT is extracted from the MR imaging; Optionally, the three-classification is obtained by comparing the imaging features with the preset thresholds of the amyotrophic stage, the frozen stage, and the thawing stage for comparison and judgment; Optionally, the three-category classification is replaced with: obtaining a classification result by classifying using a classification model trained with image feature data of the amyotrophic stage, the freezing stage, and the thawing stage; Optionally, the method further includes clinical data. Clinical data is obtained, and a classification result is obtained based on the clinical data and the image features. The clinical data includes one or more of the following: Constant-Murley score, visual analogue score.

8. A computer program product, which includes a computer program or instructions thereon, characterized in that The computer program or instruction is executed by a processor to implement the prediction method for the clinical staging of adhesive capsulitis according to any one of claims 1-7.

9. A computer device, comprising a memory, a processor, and a computer program or instruction stored on the memory, characterized in that, The computer program or instruction is executed by a processor to implement the prediction method for the clinical staging of adhesive capsulitis according to any one of claims 1-7.

10. A computer-readable storage medium having a computer program or instructions stored thereon, characterized in that, The computer program or instruction is executed by a processor to implement the prediction method for the clinical staging of adhesive capsulitis according to any one of claims 1-7.

Images